Expression and signaling of group I metabotropic glutamate receptors in astrocytes and microglia.

Stimulation of astrocytes with the excitatory neurotransmitter glutamate leads to the formation of inositol 1,4,5-trisphosphate and the subsequent increase of intracellular calcium content. Astrocytes express both ionotropic receptors and metabotropic glutamate (mGlu) receptors, of which mGlu5 receptors are probably involved in glutamate-induced calcium signaling. The mGlu5 receptor occurs as two splice variants, mGlu5a and mGlu5b, but it was hitherto unknown which splice variant is responsible for the glutamate-induced effects in astrocytes. We report here that both mRNAs encoding mGlu5 receptor splice variants are expressed by cultured astrocytes. The expression of mGlu5a receptor mRNA is much stronger than that of mGlu5b receptor mRNA in these cells. In situ hybridization experiments reveal neuronal expression of mGlu5b receptor mRNA in adult rat forebrain but a strong neuronal expression of mGlu5a mRNA only in olfactory bulb. Signals for mGlu5a receptor mRNA in the rest of the brain were diffuse and weak but consistently above background. Activation of mGlu5 receptors in astrocytes yields increases in inositol phosphate production and transient calcium responses. It is surprising that the rank order of agonist potency [quisqualate > (2S,1 'S,2'S)-2-(carboxycyclopropyl)glycine = trans-(1S,3R)-1-amino-1,3-cyclopentanedicarboxylic acid (1S,3R-ACPD) > glutamate] differs from that reported for recombinantly expressed mGlu5a receptors. The expression of mGlu5a receptor mRNA and the occurrence of 1S,3R-ACPD-induced calcium signaling were found also in cultured microglia, indicating for the first time expression of mGlu5a receptors in these macrophage-like cells.

Expression of metabotropic glutamate receptor subtype mRNA (mGluR1-8) in human cerebellum.

The expression of metabotropic glutamate receptor (mGluR) mRNAs in the human cerebellar neocortex was investigated using in situ hybridization. mGluR1, 3 and 4 mRNAs were most abundant and widely distributed, whereas mGluR2, 5 and 7 mRNA expression was circumscript. mRNAs coding for mGluR1, 3, 4 and 7 were expressed in Purkinje cells. mGluR1 and 4 mRNAs detectable in granule cells and mGluR1, 2, 3, 4 and 7 mRNAs in Golgi cells. mGluR5 mRNA was detectable in putative Bergmann glia as well as mGluR3 mRNA, which was widely expressed in glial cells. mGluR8 mRNA expression was very low in a subset of stellate/basket cells of the molecular layer whereas mGluR6 mRNA was not detectable in the cerebellar cortex.

Functional coupling of human metabotropic glutamate receptor hmGlu1d: comparison to splice variants hmGlu1a and -1b.

Functional coupling of the human mGlu1 splice variants was examined by heterologous expression. In cells stably (CHO) or transiently (A9) expressing the hmGlu1d receptor. agonists elevated intracellular calcium with a rank order of potency typical of a group I mGlu receptor (quisqualate > L-glutamate > (S)-dihydroxyphenylglycine > (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid (1S,3R-ACPD)). These responses were reduced by the antagonist (+)-alpha-methyl-4-carboxyphenylglycine (MCPG), by pretreatment with pertussis toxin and phorbol ester, and by removal of extracellular calcium. In transiently transfected HEK293 cells, the hmGlu1b and -1d receptors increased inositol monophosphate (IP) production only in the presence of glutamate, whereas hmGlu1a coupled even in the absence of agonist. This was not due to differences in receptor expression levels as assessed by immunoblotting. Adenylate cyclase activity in HEK293 cells expressing the hmGlu1 variants was neither stimulated nor inhibited by glutamate. In A9 cells hmGlu1a-mediated calcium/fluo-3 fluorescence was sensitive to depletion of intracellular calcium stores by thapsigargin, but the hmGlu1d response was resistant. Thus, hmGlu1d receptors can be distinguished from hmGlu1a by their lack of agonist-independent coupling and their dependence on extracellular calcium.

Differential expression of rat and human type I metabotropic glutamate receptor splice variant messenger RNAs.

The type I metabotropic glutamate receptor (mGlu1) messenger RNA and protein are known to be widely expressed in rat brain, but knowledge of the regional expression of splice variants other than mGlu1a is limited. Probes were designed for in situ hybridization that specifically recognize each of the carboxy-terminal splice variants mGlu1a, -1b, -1c and -1d. The novel rat mGlu1d sequence was obtained by polymerase chain reaction and the predicted protein is highly homologous to the human sequence but contains both conservative and radical substitutions and is slightly longer (912 vs 908 amino acids). Each rat mGlu1 splice variant messenger RNA was found in a unique expression pattern. The messenger RNA encoding mGlu1a was abundant in cerebellar Purkinje cells and in mitral and tufted cells of the olfactory bulb. Strong expression was also detected in hippocampal interneurons, and neurons of the thalamus and substantia nigra, while moderate expression was found in colliculi and cerebellar granule cells. The mGlu1b messenger RNA was strongly expressed in Purkinje cells, hippocampal pyramidal neurons, dentate gyrus granule cells and lateral septum, and moderately expressed in striatal, superficial cortical and cerebellar granule neurons. The mGlu1d messenger RNA was expressed in all regions where mGlu1a and -1b were detected; abundant in Purkinje cells, mitral and tufted cells, and hippocampal principal neurons and interneurons, strong in thalamus and substantia nigra, and moderate in lateral septum, cortex, striatum and colliculi. Human mGlu1 splice variant expression in the cerebellum matched that found for the rat. No specific signal was found with a probe capable of hybridizing to the rat mGlu1c splice junction, although another probe designed against a more 3' sequence of mGlu1c gave strong signals in the cerebellum and hippocampus, and moderate signals in thalamus and colliculi. It is concluded that mGlu1d messenger RNA is widely expressed, that mGlu1a and -1b messenger RNAs are expressed in almost complementary patterns and that formation of the mGlu1c splice junction is a rare event.

Benzodiazepine dependence: from neural circuits to gene expression.

The neural mechanisms underlying benzodiazepine dependence remain equivocal. The present studies tested the hypothesis that similar neural systems are recruited during diazepam tolerance and withdrawal, and that these are associated with changes in GABA(A) receptor properties. 2-Deoxyglucose quantitative autoradiography was employed to map the brain structures affected during chronic treatment and withdrawal from diazepam (5 mg/kg i.p. daily) in rats. Acute administration of diazepam evoked widespread reductions in local rates of cerebral glucose (LCGU) utilization throughout the brain. Brain structures associated with sensory processing developed tolerance to these depressant effects of diazepam after 3 days of treatment, whereas tolerance occurred in the Papez circuit of emotion after 28 days of treatment. These data suggest that adaptive changes in different neuroanatomical circuits may underlie tolerance to the various effects of diazepam. During flumazenil-precipitated withdrawal from diazepam there were marked increases in glucose use in structures of the Papez circuit, the nucleus accumbens, and the basolateral amygdala. These data suggest that the Papez circuit features strongly in diazepam tolerance and withdrawal and supports a common adaptive process being involved in these phenomena. While GABA enhancement of benzodiazepine binding was reduced in the nucleus accumbens after repeated diazepam treatment, there was little evidence to support adaptive changes in GABA(A) receptors or GABA(A) subunit gene expression (gamma2, alpha1, or alpha4) as underlying the functional changes in the identified circuits. Alternative neurochemical mechanisms, such as changes in glutamatergic function should be considered.

Enhanced expression of metabotropic glutamate receptor 3 messenger RNA in the rat spinal cord during ultraviolet irradiation induced peripheral inflammation.

Metabotropic glutamate receptors are thought to play a role in the development and maintenance of spinal hyperexcitability resulting in hyperalgesia and pain. In this study we have used in situ hybridization to investigate the distribution of metabotropic glutamate receptors mGluR1-7 messenger RNA in the rat spinal cord in a model of inflammatory hyperalgesia. Hyperalgesia was induced in nine-day-old rats by exposure of the left hindpaw to an ultraviolet light source. Lumbar portions of spinal cords were removed from control and ultraviolet-treated animals. In situ hybridization with specific oligonucleotide probes was used to localize metabotropic glutamate receptor messenger RNAs. mGluR1, 3-5 and 7 subtype messenger RNA was detected in the gray matter of the spinal cord with distribution being specific for the different subtypes. A significant increase in the expression of mGluR3 messenger RNA was seen in cells of the dorsal laminae in both sides of the lumbar spinal cord. This increase was most pronounced in laminae II, III and IV but gradually decreased and disappeared by the third day of inflammation. In parallel with this, behavioural experiments revealed mechanical hyperalgesia in both hindlimbs after ultraviolet irradiation. There was no change in mGluR3 messenger RNA expression in the thoracic segments. No changes have been detected in the levels of expression of mGluR 1,2,4,5,7 subtype messenger RNA in spinal cords taken from hyperalgesic animals. These observations show that during ultraviolet irradiation induced inflammation, the synthesis of mGluR3 messenger RNA is altered suggesting that regulation of metabotropic glutamate receptor expression may be instrumental in plastic changes within the spinal cord during the development of hyperalgesia and pain.

Behavioral sensitivity and ethanol potentiation of the N-methyl-D-aspartate receptor antagonist MK-801 in a rat line selected for high ethanol sensitivity.

The role of the N-methyl-D-aspartate (NMDA) receptors in differential ethanol sensitivity of the alcohol-insensitive [alcohol-tolerant (AT)] and alcohol-sensitive [alcohol-nontolerant (ANT)] rat lines selected for low and high sensitivity to ethanol-induced (2 g/kg) motor impairment was studied in behavioral and neurochemical experiments. A noncompetitive antagonist of the NMDA receptor, dizocilpine maleate (MK-801; 0.2 mg/kg), impaired motor function in ANT rats, but not in AT rats, in a tilting plane test. The impairment was further potentiated by a dose (0.75 g/kg) of ethanol, which alone was inactive. This effect was apparently not associated with the locomotor stimulation produced by MK-801 (0.1 and 0.2 mg/kg), because stimulation did not differ between the rat lines. Locomotor stimulation was potentiated by the low ethanol dose in both rat lines. Ethanol treatment decreased the cerebellar and hippocampal cGMP concentrations both with and without MK-801 pretreatment in both rat lines. In situ hybridization using oligonucleotide probes specific for NMDA receptor subunit mRNAs NR1 and NR2A, B, C, and D revealed no clear differences in brain regional expression between ANT and AT rates. These results indicate that the alcohol-sensitive ANT rats are very sensitive to a low dose of ethanol in the presence of NMDA receptor antagonism, consistent with the hypothesis that this receptor system is involved in acute ethanol intoxication.

Cloning, distribution and functional expression of the human mGlu6 metabotropic glutamate receptor.

The cDNA encoding the human metabotropic glutamate receptor type 6 (hmGlu6) was isolated from a human retinal cDNA library. The deduced primary sequence (877 amino acids) of the hmGlu6 receptor was 93.5% identical to its rat counterpart and shared 69.8% sequence identity with the related hmGlu4 receptor clone (912 amino acids), isolated in parallel from a human brain cDNA library. In situ hybridization revealed that the hmGlu6 mRNA is highly expressed in cells located in the inner nuclear layer of the human retina, presumably bipolar neurons. Neither PCR analysis nor in situ hybridization could detect hmGlu6 mRNA in human brain. When stably expressed in Chinese hamster ovary cells (CHO-K1) the hmGlu6 receptor inhibited adenylate cyclase through a pertussis toxin-sensitive G-protein, and reduced forskolin-elevated cyclic adenosine monophosphate (cAMP) levels in response to agonists. The rank order of agonist potency was L(+)-2-amino-4-phosphonobutyric acid (L-AP4) > L-serine-O-phosphate > L-glutamate > quisqualate = (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid ((1S,3R)-ACPD). (2S,3S,4S)-alpha-(carboxycyclopropyl)-glycine (L-CCG-I) was a partial agonist at the hmGlu6 receptor, with a potency approaching that of L-serine-O-phosphate.

Regional, developmental and interspecies expression of the four NMDAR2 subunits, examined using monoclonal antibodies.

Mouse monoclonal antibodies were raised against bacterially expressed protein sequences of the NR2A, NR2B, NR2C and NR2D subunits of the rat NMDA receptor. From immunoblots of rat brain proteins, the apparent molecular weights of these subunits were 165, 170, 135 and 145 kDa, respectively. Proteins of similar masses were observed on immunoblots of specifically transfected HEK293 cells. Deglycosylation with endoglycosidase F reduced the mass of each endogenous NR2 subunit by approximately 10 kDa. In distribution studies, NR2A-immunoreactive protein (IRP) was located throughout the adult rat brain, NR2B-IRP was primarily in the forebrain, NR2C-IRP was predominantly in the cerebellum and NR2D-IRP was mainly found in the thalamus, midbrain and brainstem. Whereas NR2A- and NR2C-IRPs increased during rat brain post-natal development, NR2B- and NR2D-IRPs were abundant at birth and declined with age, especially in cerebellum. NR2-IRPs of mouse, rabbit, frog and human brain were of sizes similar to those of the corresponding rat subunits and were similarly distributed. In summary, NR2 subunits are large glycoproteins whose specific expression profiles in the brain are developmentally and regionally regulated and which are similarly expressed in a variety of species.

Erratum to 'HmGlu1d, a novel splice variant of the human type I metabotropic glutamate receptor' [Eur. J. pharmacol. 296 (1996) R1-R3].

A novel splice variant, hmGlu1d, of the human mGlu1 metabotropic glutamate receptor has been isolated from a human brain library. This clone is identical to human mGlu1a except that it lacks 35 nucleotides in the 3' coding sequence, which predicts a truncated protein and a novel carboxy terminus. After injection of the encoding sequence into mouse A9 fibroblasts, quisqualate and (1S, 3R)-aminocyclopentane-1, 3-dicarboxylic acid ((1S,3R)-ACPD) elicited concentration-dependent increases in intracellular Ca2+ (pEC50 values of 6.09 and 4.33, respectively).

HmGlu1d, a novel splice variant of the human type I metabotropic glutamate receptor.

A novel splice variant, hmGlu1d, of the human mGlu1 metabotropic glutamate receptor has been isolated from a human brain library. This clone is identical to human mGlu1a except that it lacks 35 nucleotides in the 3' coding sequence, which predicts a truncated protein and a novel carboxy terminus. After injection of the encoding sequence into mouse A9 fibroblasts, quisqualate and (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid ((1S,3R)-ACPD) elicited concentration-dependent increases in intracellular Ca2+ (pEC50 values of 6.09 and 4.33, respectively).

The distribution of splice variants of the NMDAR1 subunit mRNA in adult rat brain.

Regional variation in the alternative splice forms of the NMDAR1 subunit mRNA was investigated by in situ hybridization in the adult rat brain, using radiolabelled splice-specific oligonucleotide probes. Each splice variant was detected in an individual distribution. The NMDAR1-a and NMDAR1-2 forms were widely and abundantly distributed throughout the brain, except for the inferior colliculus. The NMDAR1-b and NMDAR1-4 variants were located in similar patterns in fewer areas (e.g. parietal cortex, hippocampus CA3, thalamus, inferior colliculus, cerebellar granule cells). In contrast, the NMDAR1-1 forms were distributed in a pattern approximately complementary in the forebrain to that of NMDAR1-4 (weakly expressed in thalamus and inferior colliculus). The NMDAR1-3 variants were not abundant in any structure. Considerable overlap of the in situ hybridization images was noted, so all eight splice combinations are possible in heterogenous distributions. Correlation of the distribution of NMDAR1 mRNA splice forms with functional analyses of heteromeric recombinant receptors will be necessary to ascertain if alternative splicing of the NMDAR1 subunit can account for some of the known heterogeneity of endogenous NMDA receptors.

Cellular and subcellular distribution of NMDAR1 splice variant mRNA in the rat lumbar spinal cord.

The regional distribution of alternatively spliced messenger RNA encoding the N-methyl-D-aspartate (NMDA) receptor R1 subunit (NMDAR1) variants was examined by in situ hybridization in the rat lumbar spinal cord. Splice-specific oligonucleotide probes [recognizing full-length mRNA (NMDAR1-1), deletion exon 21 (NMDAR1-2), deletion exon 22 (NMDAR1-3), combined deletion exons 21 and 22 (NMDAR1-4) and mRNA which lacks (NMDAR1-a) or contains exon 5 (NMDAR1-b)] detected marked differences in abundance and distribution of N- and C-terminal spliced variants. The NMDAR1-a, NMDAR1-2 and NMDAR1-4 mRNAs were evenly distributed throughout all laminae of the dorsal and ventral horns. In the superficial dorsal horn NMDAR1-b mRNA was preferentially detected in laminae II inner and III, while NMDAR1-1 mRNA was restricted to laminae I to III. Large neurons in laminae IV and V contained mainly NMDAR1-a, NMDAR1-2 and NMDAR1-4 mRNAs and occasionally NMDAR1-b. The NMDAR1-3 variant was only detected in very low abundance, being restricted to occasional cells in lamina I and II. In the ventral horn, motor neurons showed strong signals for NMDAR1-a, NMDAR1-b, NMDAR1-2 and NMDAR1-4 mRNAs. Serial sectioning through large motor neurons permitted the detection of multiple splice variants in single neurons. Analysis of the subcellular distribution of the mRNAs revealed that the NMDAR1-1 mRNA was almost exclusively found in the cell nucleus, NMDAR1-a mRNA was largely in the cytoplasm, while all other splice variants showed a homogeneous distribution between nucleus and cytoplasm. Comparison of the in situ hybridization images with functional analyses of heteromeric recombinant receptors will be necessary to ascertain whether splice variants of the NMDAR1 receptor subunit can account for some of the known electrophysiological properties of spinal cord neurons under physiological and pathophysiological conditions.

Ligand binding profile of the rat metabotropic glutamate receptor mGluR3 expressed in a transfected cell line.

A cDNA clone encoding the rat metabotropic glutamate receptor mGluR3 was stably transfected into human embryonic kidney 293 cells. Receptor-expressing cell lines were characterized by centrifugation binding assays using [3H]glutamate as radioligand. The rank order of affinity was L-glutamate > (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid (1S,3R-ACPD) > L(+)-2-amino-3-phosphonopropionic acid (L-AP3) > quisqualic acid > L(+)-2-amino-4-phosphonobutyric acid (L-AP4) > ibotenic acid. The active enantiomers of several phenylglycines displayed Ki values of 300 to 400 microM. The nonactive enantiomers and the standard ionotropic glutamate receptor ligands N-methyl-D-aspartic acid (NMDA), (R,S)-alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and kainic acid only weakly displaced [3H]glutamate. In this cell line, L-glutamate and (2S,3S,4S)-alpha-(Carboxycyclopropyl)-glycine (L-CCG-I) reduced cAMP levels in a dose-dependent manner. The sensitivity of this system and its easy applicability make it feasible to envisage ligand binding assays on cell lines expressing cloned receptors as useful screening tools to discover and characterize new and specific agonists and antagonists.

Localization and developmental expression of the NMDA receptor subunit NR2A in the mammalian retina.

The localization of the N-methyl-D-aspartate receptor subunit NR2A was studied, by using light microscopic immunocytochemistry, in the retina of adult rat, rabbit, cat, and monkey. Strong, punctate immunolabeling was observed in the inner plexiform layer indicating a synaptic localization of the NR2A subunit. The punctate labeling was concentrated in two bands corresponding to the on- and off-sublaminae of the inner plexiform layer. The punctate character of immunofluorescence suggested a synaptic localization of the receptor. This was confirmed by electron microscopy of immunostained adult rat retina. The staining was localized postsynaptic to cone bipolar cells, and only one of the two postsynaptic elements of the dyad was labeled. Staining was not observed at extrasynaptic plasma membranes. In situ hybridization of adult rat retina showed expression of the NR2A subunit in virtually all ganglion cells and displaced amacrine cells in the ganglion cell layer and in a subset of amacrine cells in the inner nuclear layer. The postnatal developmental expression of the NR2A subunit was studied in rat retina by light microscopic immunocytochemistry. Punctate immunolabeling appeared prior to eye opening, and the developmental profile of NR2A could be compatible with a role in development of circuitry in the inner plexiform layer.

Ligand affinities at recombinant N-methyl-D-aspartate receptors depend on subunit composition.

The ligand preferences of recombinant NR1 homomeric and NR1-NR2 heteromeric NMDA receptors were examined by homogenate binding assay. The binding affinities for most ligands were similar to those reported for native NMDA receptors. The order of affinity for [3H]glutamate was NR1-NR2B > NR1-NR2A approximately NR1-NR2D > NR1-NR2C > NR1. NMDA had approximately equal affinity for all heteromeric types (Ki approximately 5 microM), but the competitive antagonists CGS 19755 (cis-4-(phosphonomethyl)piperidine-2-carboxylic acid) and CGP 39653 (D,L-(E)-2-amino-4-propyl-5-phosphono-3-pentenoic acid) displayed the affinity order NR1-NR2A > NR1-NR2B > NR1-NR2D > NR1-NR2C. Binding of [3H]CGP 39653 could only be detected at the NR1-NR2A receptor type (Kd approximately 6 nM). The glycine site antagonist [3H]5,7-dichlorokynurenate bound with good affinity to all recombinant receptors (Kd approximately 50-100 nM), while glycine exhibited an affinity order of NR1-NR2C >> NR1 = NR1-NR2B = NR1-NR2D > NR1-NR2A. The channel-site ligand [3H]MK 801 ((+)-5-methyl-10,11-dihydro-5H- dibenzo[a,d]cyclo-hepten-5,10-imine hydrogen maleate) showed the affinity ranking NR1-NR2A = NR1-NR2B >> NR1 > NR1-NR2C = NR1-NR2D. Thus the ligand binding affinities of recombinant NMDA receptors is dependent on their subunit composition. The NR1-NR2A, NR1-NR2B, NR1-NR2C and NR1-NR2D receptors may account for the antagonist-preferring, agonist-preferring, cerebellar, and medial thalamic subtypes of native NMDA receptors, respectively.

Regional and developmental heterogeneity in splicing of the rat brain NMDAR1 mRNA.

Developmental and regional alternative splicing of the NMDAR1 subunit gene transcript was examined by in situ hybridization in the developing and adult rat brain. NMDAR1 mRNA, barely detectable at embryonic day 14, increased gradually during development until the third postnatal week, after which it declined slightly to adult levels, when it was detected in every examined neuronal type. Each splice form of the primary NMDAR1 gene transcript was found to follow a parallel profile of abundance in the brain, but marked regional differences were observed in splicing at both 5' and 3' sequences. The individual regional distributions of splice forms appeared to be established around birth, with little change thereafter, except in the overall abundance. The NMDAR1-a and NMDAR1-2 splice forms occurred extensively and approximately homogeneously throughout brain gray matter. The NMDAR1-b variant was found primarily in the sensorimotor cortex, neonatal lateral caudate, thalamus, hippocampal CA3 field, and cerebellar granule cells, but was absent from adult caudate. The NMDAR1-1 and -4 splice forms were detected in almost complementary patterns; the former was concentrated in more rostral structures such as cortex, caudate, and hippocampus, while the latter was principally in more caudal regions such as thalamus, colliculi, and cerebellum. These two splice forms accounted for a greater proportion of the adult NMDAR1 mRNA than that of the neonate. The NMDAR1-3 mRNA variant was scarce, being detected only at very low levels in postnatal cortex and hippocampus. The different splice forms may generate regional differences in NMDA receptor properties during development and in the adult CNS.

Developmental and regional expression in the rat brain and functional properties of four NMDA receptors.

An in situ study of mRNAs encoding NMDA receptor subunits in the developing rat CNS revealed that, at all stages, the NR1 gene is expressed in virtually all neurons, whereas the four NR2 transcripts display distinct expression patterns. NR2B and NR2D mRNAs occur prenatally, whereas NR2A and NR2C mRNAs are first detected near birth. All transcripts except NR2D peak around P20. NR2D mRNA, present mainly in midbrain structures, peaks around P7 and thereafter decreases to adult levels. Postnatally, NR2B and NR2C transcript levels change in opposite directions in the cerebellar internal granule cell layer. In the adult hippocampus, NR2A and NR2B mRNAs are prominent in CA1 and CA3 pyramidal cells, but NR2C and NR2D mRNAs occur in different subsets of interneurons. Recombinant binary NR1-NR2 channels show comparable Ca2+ permeabilities, but marked differences in voltage-dependent Mg2+ block and in offset decay time constants. Thus, the distinct expression profiles and functional properties of NR2 subunits provide a basis for NMDA channel heterogeneity in the brain.

Flumazenil induces localised increases in glucose utilization during diazepam withdrawal in rats.

The quantitative [14C]2-deoxyglucose autoradiographic technique has been employed to identify the neural circuits involved in diazepam withdrawal. Local cerebral glucose utilization (LCGU) was assessed in parallel groups of rats chronically treated with diazepam (5 mg/kg i.p., daily for 28 days), in rats that were withdrawn from chronic diazepam 24 h previously and in those that received flumazenil (5 mg/kg i.v.) immediately or 24 h after the last dose of diazepam. Two further groups received chronic vehicle or acute flumazenil (5 mg/kg i.v.). Rats withdrawn from diazepam 24 h previously did not produce changes in LCGU in the 51 structures examined compared with both control and chronic diazepam treated groups, suggesting that spontaneous withdrawal from small doses of diazepam does not evoke marked alterations in functional activity. In contrast, flumazenil-precipitated diazepam withdrawal produced a marked increase in glucose use in structures of the Papez circuit of emotion (mammillary body, anterior thalamus, cingulate cortex), together with increases in the septal nucleus, basolateral amygdala and nucleus accumbens. Less widespread increases in glucose use occurred in primary auditory and visual areas and in extrapyramidal areas. This pattern resembles that produced after acute FG-7142 administration (Brain Res., 475 (1988) 218-231). In rats receiving flumazenil 24 h after the last dose of diazepam there was a similar, but more restricted, pattern of change in LCGU. Flumazenil had no effect on LCGU in drug naive rats. Thus, flumazenil could only exert an effect upon LCGU in rats chronically treated with diazepam. These data provide functional neuroanatomical evidence for a withdrawal shift in the inverse agonist direction after chronic diazepam and suggest that flumazenil-precipitated withdrawal changes may merely be a reflection of this phenomenon.